1. Chapter 12 Application of Changing Criterion Design

2. Basic Changing Criterion Design: Increasing Behavior
Research Question: The authors demonstrated the effectiveness of a graphic feedback procedure to increase rate of breakfast tray stripping (i.e., cleaning) for a young woman with moderate mental retardation.
Subjects: The subject was a 20-year old woman with moderate mental retardation (IQ of 42 on Wechsler Intelligence Scale for Children) who had resided at home all her life.
Setting: The study took place in a hospital kitchen where dishes were washed in a production line format.

3. Dependent Variable: The dependent variable was the speed (rate per minute) of stripping trays.
Stripping trays consisted of disposing of waste, stacking dishes, and soaking silverware.
Also, a corollary dependent measure (questionnaire) was used to determine how the subject’s coworkers perceived the effects of the intervention. This was administered to the kitchen supervisor and consisted of eight questions concerning goal, procedures, and outcomes of the study.
Independent Variable: The independent variable was graphic feedback concerning rate of work for the first three intervention subphases.
During the fourth subphase, verbal feedback for sloppiness was added to graphic feedback. In the fifth and final subphase, only graphic feedback was provided. A changing criterion was also used.

4. The Design
The subject’s rate of tray stripping was evaluated using a changing criterion design to produce an increase in the target behavior.
Criteria during the intervention phase were changed when met for at least 3 consecutive days, with one exception. If the subject’s performance was increasing, the criterion was not changed until performance stabilized or a decreasing trend was evident.

5. Intervention
The subject’s primary task was breakfast tray stripping
Each training session lasted the duration of time required to strip 130 to 150 trays. Based on the subject’s rate for 16 baseline observations and the competitive rate, criteria for the various subphases were determined.
Initial criteria were set based on baseline performance.
The trainer timed one regular tray stripper and determined a rate of 2.38 trays per minute was commensurate with a competitive employment rate.
As a result, 2.4 trays per minute became her criterion from the third intervention subphase to the end of the study.
The questionnaire was completed by all coworkers within 1 month of the study’s termination.

6. Obtaining the Data and Plotting the Results
The rate of tray stripping was recorded daily.
The mean was calculated by measuring the length of time it took the subject to strip 10 consecutive randomly selected trays.
The total number of minutes to strip the trays was divided by 10 to determine the mean number of trays per minute.
Interobserver reliability mean was .999, with no measure below .997.

7. Results
During Phase 1 (baseline), the subject stripped 1.33 trays per minute on average across all baseline observations.
During the first intervention phase, when tray stripping Criterion 1 (1.8 trays per minute) and graphic feedback were in effect, her production rate increased to an average of 1.9 trays per minute.
During Criterion 2, the subject’s tray stripping mean performance was 2.05 trays per minute.
During Criterion 3, the subject’s tray stripping speed increased but her coworkers did not accept the quality of her work. As a result, Criterion 3 was extended for another 9 days and the subject was provided with verbal feedback on the quality of her work.

8. Results
Following the feedback, her quality of work improved and it was accepted by her coworkers.
Verbal feedback was then dropped from the intervention.
From the beginning to the end of the program, the subject’s tray stripping rate increased represented a savings of 44 to 51 minutes per day over baseline levels.
The results, of the questionnaire indicated that the majority of coworkers believed the intervention was acceptable and was responsible for the subject’s improved performance.

9. Why Use a Changing Criterion Design?
A changing criterion design was very appropriate because it allowed the trainer to change the criterion gradually and demonstrate that increases in tray stripping were due to the changing criterion design and the feedback given to the subject.

10. Limitations of the Study
There are two limitations:
First, the training program ended while the subject was performing at an acceptable level but the trainer was present.
To ensure that the subject would maintain competitive work rates, the trainer should have gradually faded his or her presence and feedback in order to bring the subject’s work performance under natural conditions.
Second, the trainer did not allow the subject to strip lunch or dinner trays to generalize her skills.

11. Basic Changing Criterion Design: Decreasing Behavior.
Research Question: The researchers were interested in determining the effectiveness of a behavioral treatment program in reducing the ingestion of caffeine.
Subjects: Three adults subjects volunteered for the study. They each:
drank eight or more cups of coffee daily (over 1,000 mg caffeine)
reported physiological and behavioral symptoms (e.g., nervousness)
wanted to decrease their caffeine intake

12. Dependent variable: The total daily caffeine intake.
Setting: The subjects self-monitored their coffee drinking throughout the day in any and all environments.
Dependent variable: The total daily caffeine intake.
Independent Variable: A treatment program used effectively to reduce smoking was used.
The Design: Basic Changing Criterion Design for decreasing behavior.
This included a baseline, four treatment phases, and a follow-up phase that included probes up to 40 weeks after the program ended.

13. Intervention
Each subject deposited $20.00 at the onset of the program, which would be reimbursed if they met the target goals.
They could earn $10.00 back in four payments of $2.50 if they met the four criterion level specified by the researchers.
If they did not exceed the criterion level on any day within a treatment phase, they would also receive a $1.00 bonus and be eligible to receive a $10.00 bonus at the end of the study.
If they exceeded the criterion level within any treatment phase they would lose the $2.50, the bonus, and the eligibility.

14. Obtaining the Data and Plotting the Results
Subjects self-monitored their caffeine intake by recording the number and type of caffeine drinks they consumed.
They used a provided conversion chart to keep a running record of their daily caffeine intake. At the end of the day, they recorded their total intake.
Determined from baseline data, four criterion levels were specified that would gradually reduce the caffeine intake to 600 mg or less.
After the treatment phases were completed, each subject was called every 2 weeks and asked to monitor his or her intake for the next 2 days.

15. Results
Subject 1 decreased her caffeine intake from an average of 1,008 mg per day during baseline to an average of 357 mg per day during final treatment phase. Her intake at the 40-week follow-up was further reduced to slightly under 300 mg.
Subject 2 decreased her caffeine intake from an average of 1,147 mg at baseline to an average of 357 mg during the final treatment phase to an average of 250 mg during follow-up.
Subject 3 decreased his caffeine intake from an average of 1,175 mg during baseline to an average of 314 mg during final treatment phase. However, he exceeded the criterion level for 1 day in both the third and fourth treatment phases. His consumption level during follow-up increased to an average of 537 mg.

16. Why Use a Changing Criterion Design?
The nature of this study makes it ideal for a changing criterion design.
The target behavior, consumption of caffeine beverages, can be decreased in a gradual, stepwise fashion.
The visual nature of the graphs aided the self-monitoring procedure.

17. Limitations of the Study
Although positive results were found in this study, it is difficult to determine which aspects of the program was responsible.
The authors also noted that other factors, such as the subjects’ desire to reduce their consumption and social pressure, could have also affected the results.
The reliability of self-monitoring procedure is also a concern.

18. Changing Criterion Design with Treatment Phases of Different Lengths
Research Question: The investigators demonstrated the efficacy of an intervention program to enhance recognition of commonly used similes by hearing-impaired students.
Subjects: Two 20-year old college students.
Subject 1 was male who had congenital bilateral severe to profound sensorineural hearing loss with etiology unknown.
Subject 2 was a female full-time resident client with bilateral severe sensorineural hearing loss due to maternal rubella.
*Both subjects wore hearing aids during the investigation.

19. Setting: Subjects were enrolled in an adult therapy program at a university speech and hearing clinic. Therapy was conducted in a room equipped with a microphone and a two-way mirror leading to an adjoining observation booth.
Dependent Variable: The subjects’ correct application of eight similes (e.g., “As happy as a lark”).
The authors developed 18 incomplete passages that would be relevant for each of the eight similes.

20. Independent Variable: The independent variable was the intervention program involving instruction, review, exemplars, and feedback along with a changing criterion for reinforcement.
The Design: The investigators used a changing criterion design with treatment phases of different lengths to evaluate the treatment effects.

21. At the first session of baseline observations, subjects:
Intervention
At the first session of baseline observations, subjects:
a) read the definition of the term simile
b) were given examples of similes
c) were asked to match each of 144 passages to one of eight correct similes
During the subsequent baseline observations, subjects were given 24 opportunities to respond to randomly selected passages by stating a correct simile.

22. Criteria was set based on the following:
Intervention (cont.)
There were eight intervention subphases corresponding to the eight target similes.
During the intervention sessions, subjects were given specific reinforcement for correct responses and correction procedures for incorrect ones.
Sessions were conducted three times a week for 30 minutes per session for a total of 35 sessions.
Criteria was set based on the following:
Subjects should recognize at least three additional novel passages of the 24 presented.
Accuracy should be maintained for previously developed skills at 80% or higher.

23. Obtaining the Data and Plotting the Results
Two independent observers recorded the number of correct and incorrect applications of the eight similes to the 24 incomplete passages selected from the original 144.
The researchers only reported the total number for each session (maximum of 24) since the number of opportunities was held constant.
Assessment data were graphed and shared with the subjects.
Interobserver reliability was obtained at least once in each subphase by two independent observers.

24. During treatment sessions:
Results
Subjects 1 and 2 did not correctly identify any of the similes during the baseline sessions.
During treatment sessions:
Subject 1’s criterion level for the first seven intervention subphases were 2, 5, 9, 1, 13, 17, and 20 (out of a possible 24. During subphase 8, he obtained a correct response rate of 23 in each of the last three sessions.
Subject 2 completed only seven subphases of the program due to time constraints. The criterion levels that were set were 2, 4, 7, 11, 14, 18, and 22. During subphase 7, she obtained correct response rates of 20, 22, 24, and 23 in the last four sessions.

25. Why Use a Changing Criterion Design with Treatment Phases of Different Lengths?
When the goal is a gradual increase in student’s performance.
It provides a logical, systematic, and positive evaluation of the intervention package designed to increase simile recognition skills of these subjects.
Allows for modest expectations for performance at the beginning and then increases as the students become more proficient.

26. Limitations of the Study
The changing criterion design study required high levels of involvement in collecting data and providing systematic instruction.
Without support from others, a single therapist may have found it would be impossible to investigate a program of this nature.

27. Changing Criterion Design with A Return to Baseline Phase
Research Question: To determine if variable ratio (VR) schedules would increase and maintain high rates of exercise.
Subjects: The subjects were six boys, each 11 years old. Three of these subjects were obese and three of them were of normal weight.
Setting: The setting was conducted in an elementary school nurse’s room. Each subject exercised on a stationary bicycle in the nurse’s room.

28. Dependent Variable: The target behavior was pedaling on a stationary bicycle.
Pedaling was measured in the subjects’ overall responding rate per session and the total time spent exercising per minute.
Independent Variable: Token system in which subjects earned points by pedaling bicycle in exchange for reinforcers from a reinforcement menu.
Menu included a handheld battery-operated game, kite, bicycle bell, flashlight, model car, model plane, puzzle, adventure and comic books).
Different schedules and criteria for reinforcement were also used.

29. The Design: A changing criterion design with a return to baseline phase.
A VR schedule of reinforcement was implemented with a different criterion for performance of each subject.
The Intervention
Subjects were tested individually once a day, from Monday through Friday, for about 12 weeks.
After stable baseline was established, the VR schedule of reinforcement (i.e., a rate of revolutions per minute) was implemented for nine sessions.

30. The Intervention (cont.)
For each of the three VR subphases, a criterion of 15% increase over mean responding during the previous subphase was established.
In other words, a different criterion for performance was specified for each subject, based on his performance in the previous subphase.
A return to baseline phase was implemented after the third intervention subphase. The same VR subphase was reinstated.

31. Obtaining the Data and Plotting the Results
The experimenter sat behind the equipment panels out of the subject’s sight and started recording data when the subject began pedaling the bicycle.
The stationary bicycle was programmed to signal when the variable number of responses had occurred.
Each time the bell rang and illuminated the light, the subject earned a point to be used in the token system.

32. Results
During baseline, the nonobese boys responded at a mean of 71.9 revolutions per minute and the obese boys responded at a mean of 59.2 revolutions per minute.
During the first subphase, the mean rates of responding for the nonobese boys and obese boys were and revolutions per minute.
During the second subphase, the rates of responding per minute increase to and
During the third subphase, the mean response rates for the nonobese and obese boys were and
Following the third phase, a brief return to baseline produced a reduction in response rates to 95.3 for nonobese boys and 83.6 for obese boys. However, when reinforcement was reintroduce, it produced the highest response rates in all boys (nonobese boys, and obese boys,123.6).

33. Why Use Changing Criterion Design with a Return to Baseline?
This design provided the opportunity to set initial criterion based on the subjects’ own baseline performance, and allowed each subject to increase his own rate through small successive increments.
As a result, rates of exercise were increased gradually and systematically.
By reinstating the baseline conditions and noting decreases in responding in all subjects, more convincing evidence of the functional relationship between the independent and dependent variable was provided.

34. Limitations of the Study
One limitation of the study was gender bias.
It was noted in the study that there are more overweight girls than boys, yet the authors included only boys in the study.
Therefore, the findings may not be generalizable to the general population.